THEFT AND LOSSES IN TURKISH ELECTRICITY SECTOR: EMPIRICAL ANALYSIS AND IMPLICATIONS FOR TARIFF DESIGN
by
Eray Gümüşdere
Submitted to the Graduate School of Arts and Social Sciences in partial fulfillment of
the requirements for the degree of Master of Arts
in Economics
Sabanci University Summer 2004
THEFT AND LOSSES IN TURKISH ELECTRICITY SECTOR: EMPIRICAL ANALYSIS AND IMPLICATIONS FOR TARIFF DESIGN
APPROVED BY:
Assoc. Prof. Dr. İzak Atiyas ....……….
(Thesis Supervisor)
Assoc. Prof. Dr. Alpay Filiztekin ……….
Assoc. Prof. Dr. Yıldız Arıkan ……….
DATE OF APPROVAL: ……….
© Eray Gümüşdere 2004
All Rights Reserved
ABSTRACT
THEFT AND LOSSES IN TURKISH ELECTRICITY SECTOR: EMPIRICAL ANALYSIS AND IMPLICATIONS FOR TARIFF DESIGN
Since the 1980s, electricity industries in many countries have been undergoing privatization and liberalization in order to generate improvements in efficiency and quality. Liberalization and privatization can only be successful if privatized utilities can be financially viable on their own and not rely on government subsidies except for those that are designed for social objectives such as universal service. In the case of Turkey, an important impediment to privatization and liberalization are thefts and losses that generate large asymmetries in costs across different regions. In international comparisons these costs are exceptionally high and they threaten the economic and political feasibility of reform.
This master thesis undertakes an econometric analysis to identify factors that cause such large variances in electricity theft-losses across different regions of Turkey.
Once these factors are identified, empirical results are then used to derive implications for both privatizations as well as principles that should guide tariff design, including possible subsidy mechanisms.
ÖZET
TÜRK ELEKTRİK SEKTÖRÜNDE KAYIP VE KAÇAKLAR: AMPİRİK ANALİZ VE TARİFE TASARIMINA ETKİSİ
1980’lerden bu yana elektrik sektörü bir çok ülkede verimlilik ve gelişim sağlamak için özelleştirme ve liberalleşme sürecine girmiştir. Liberalleşme ve özelleştirme ancak özelleştirilmiş kurumlar merkezi hükümetten herhangi bir finansal destek almadan kendi ayakları üzerinde durabildiği zaman başarılı olabilir. Bunun tek istisnası sadece evrensel servis gibi sosyal amaçlar olabilir. Türkiye için elektrik sektöründeki serbestleşme ve özelleştirme karşısındaki en önemli engellerden biri bölgeler arasında büyük maliyet asimetrileri yaratan kayıp ve kaçaklardır. Diğer dünya devletlerine göre kayıp kaçak maliyetleri aşırı yüksektir ve bu da yapılmak istenen reformun ekonomik ve siyasi olabilirliğini tehdit etmektedir.
Bu yükseklisans tezinde bölgeler arasıdaki muazzam kayıp ve kaçak farklılıklarına neden olan faktörlerin tanımlanması için ekonometrik bir çalışma yapılmıştır. Bu faktörler tanımlandıktan sonra, ampirik çalışma sonuçlarının özelleştirme ve tarife tasarımına iması subvansiyon mekhanizması dahil olmak üzere değerlendirilmiştir.
TABLE OF CONTENTS
ABSTRACT...iv
ÖZET...v
LIST OF TABLES...viii
LIST OF FIGURES...xii
LIST OF SYMBOLS AND ABBREVIATIONS...xiii
1. INTRODUCTION AND BACKGROUND………...1
1.1. Introduction………...……...1
1.2. Process of Liberalization in the Electricity Sector………...…………2
1.3. Distribution Private Participation………...3
1.4. The Electricity Market Law……….…….5
1.4.1. Vertical Separation………...……...………6
1.4.2. Licensing……….6
1.4.3. Regulatory Authority………..………6
1.4.4. Competition in the Generation and Retail Segment……….……..6
1.4.5. Privatization………7
2. PROBLEM DESCRIPTION……….8
2.1. Definition of the Problem……….8
2.1.2. Technical Losses………..………...9
2.1.3. Theft………...………..……..11
2.1.4. Measuring Theft-losses………..………13
2.1.5. The Size of the Problem………13
2.2. Current Tariffs and Subsidy Mechanism………15
2.2.1. The Concept of Tariff………15
2.2.2. The Current Financial Flow and Cross Subsidy Mechanism…………16
3. EMPIRICAL ANALYSIS……….………..…19
3.1. The Model………...…19
3.2. Independent Variables………19
3.2.1. Economic Variables………...20
3.2.2. Social and Cultural Variables………..………..…21
3.2.3. Variables Reflecting the Enforcement Capacity and the Reach of the
State ………....22
3.2.4. Distribution Utility’s Managerial variables………...………23
3.2.5. Physical Variables………...………..…23
3.2.6. Dummy Variables……….…..………...24
3.3. The Data Set………..…….………24
3.4. Descriptive Statistics………..………....28
3.4.1. Dependent Variables………..……28
3.4.2. Independent Variables………...…32
3.5. Empirical Results………....……34
3.5.1. Year 2000 regression……….…36
3.5.2. Panel Data Analysis………...39
3.5.2.1. Multicollinearity Diagnosis………...…39
3.5.2.2. TLR Regression for the Whole Sample………41
3.5.2.3. TLR Regression for the Shrunk Sample………...…44
3.5.2.4. TLPC Regression for the Whole Sample………...…46
3.5.2.5. TLPC Regression for the Shrunk Sample……….……47
3.6. Summary and Inference………..49
4. IMPLICATIONS FOR PRIVATIZATION AND TARIFF DESIGN……....51
4.1. Privatization Concern………...51
4.2. Tariff Design Concern………54
5. CONCLUSION………...……59
APPENDIX A: THE DATA SET………..……….62
APPENDIX B: MEAN AND VARIANCE COMPARISON TESTS…………87
APPENDIX C: STATA OUTPUTS OF REGRESSIONS………..….100
BIBLIOGRAPHY………..………...127
LIST OF TABLES
Table 2.1 Cumulative Amount of Theft-Losses in MWh for 8 years) (1994- 2001………..………..…...14
Table-2.2 Methods for Tariff Regulation for Different Electricity Activities and Services………..………...….18
Table 3.1 Descriptive Statistics of Dependent Variables for the Whole
Sample………..……….….29 Table 3.2 Descriptive Statistics of Dependent Variables for the Shrunk
Sample………..……….…...29 Table 3.3 Descriptive Statistics of Independent Variables for the Whole
Sample………..………...33 Table 3.4 Descriptive Statistics of Independent Variables for the Shrunk
Sample………..………..………...…34 Table 3.5 Summary Results for Year 2000 Regressions………37 Table 3.6 Variance Inflation Factors for Year 2000 Independent Variables…..38 Table 3.7 Correlation Matrix for Independent Variables………40 Table 3.8 Principal Components for Independent Variables………..41 Table-3.9 Summary of TLR Regressions Results for the Whole Sample…..…42 Table 3.10 Summary of TLR Regressions Results for the Shrunk Sample..…..44 Table 3.11 Summary of TLPC Regressions Results for the Whole Sample..…47 Table 3.12 Summary of TLPC Regressions Results for the Shrunk Sample…..48 Table A.1 the Panel Data………63 Table A.2 Additional Data for Year 2000 Regression………..…………..84 Table B.1 Testing the Equality of Means of GDPCR for Whole and Shrunk Sample………..………...88
Table B.2 Testing the Equality of Means of RESECR for Whole and Shrunk Sample………..………...……..….88
Table B.3 Testing the Equality of Means of INDECR for Whole and Shrunk Sample………..………...89
Table B.4 Testing the Equality of Means of DEPOSITR for Whole and Shrunk Sample………..………...….89
Table B.5 Testing the Equality of Means of DEPR for Whole and Shrunk
Sample………..………..…....90 Table B.6 Testing the Equality of Means of PIE for Whole and Shrunk
Sample………..………...90 Table B.7 Testing the Equality of Means of TRRR for Whole and Shrunk Sample………..………...91
Table B.8 Testing the Equality of Means of TRPGDP for Whole and Shrunk Sample………..………...91
Table B.9 Testing the Equality of Means of PNPS for Whole and Shrunk
Sample………..………...92 Table B.10 Testing the Equality of Means of TUR for Whole and Shrunk
Sample………..………...92 Table B.11 Testing the Equality of Means of LWLPS for Whole and Shrunk Sample………..………...93
Table B.12 Testing the Equality of Means of AGRGDPR for Whole and Shrunk Sample………..……….….93 Table B.13 Testing the Equality of Variances of GDPCR for Whole and Shrunk Sample………..……….….94
Table B.14 Testing the Equality of Variances of RESECR for Whole and Shrunk Sample………..……….…94
Table B.15 Testing the Equality of Variances of INDECR for Whole and Shrunk Sample………..………..…95
Table B.16 Testing the Equality of Variances of DEPOSITR for Whole and Shrunk Sample………..………..…95
Table B.17 Testing the Equality of Variances of DEPR for Whole and Shrunk Sample………..………..…96
Table B.18 Testing the Equality of Variances of PIE for Whole and Shrunk Sample………..………..96
Table B.19 Testing the Equality of Variances of TRRR for Whole and Shrunk Sample………..……….….…97
Table B.20 Testing the Equality of Variances of TRPGDP for Whole and Shrunk Sample………..……….…97
Table B.21 Testing the Equality of Variances of PNPS for Whole and Shrunk Sample………..………..…98
Table B.22 Testing the Equality of Variances of TUR for Whole and Shrunk Sample………..………..…98
Table B.23 Testing the Equality of Variances of LWLPS for Whole and Shrunk Sample………..……….99
Table B.24 Testing the Equality of Variances of AGRGDPR for Whole and Shrunk Sample………..………..…99
Table C.1 Stata Output for Year 2000 Regression (TLR) ………101 Table C.2 Stata Output for Year 2000 Regression (TLPC) ………..…102 Table C.3 Stata Output for Between Estimator Regression using Whole Sample (TLR) ………..………...103
Table C.4 Stata Output for Fixed Effects Regression using Whole Sample (TLR)
………..……….…...104 Table C.5 Stata Output for Random Effects Regression using Whole Sample (TLR) ………..……….…105
Table C.6 Stata Output for FGLS Regression using Whole Sample (TLR)…..106 Table C.7 Stata Output for FGLS Regression with Lagged Dependent Variable using Whole Sample (TLR) ……….107
Table C.8 Stata Output for Arellano-Bond GMM Regression using Whole Sample (TLR) ………..………..……..108
Table C.9 Stata Output for Between Estimator Regression using Shrunk Sample (TLR) ………..……….109
Table C.10 Stata Output for Fixed Effects Regression using Shrunk Sample (TLR) ………..……….110
Table C.11 Stata Output for Random Effects Regression using Shrunk Sample (TLR) ………..……….111
Table C.12 Stata Output for FGLS Regression using Shrunk Sample
(TLR………..………..…….112 Table C.13 Stata Output for FGLS Regression with Lagged Dependent Variable using Shrunk Sample (TLR)……….…113
Table C.14 Stata Output for Arellano-Bond GMM Regression using Shrunk Sample (TLR) ………..………..……..114
Table C.15 Stata Output for Between Estimator Regression using Whole Sample (TLPC) ………..………...115
Table C.16 Stata Output for Fixed Effects Regression using Whole Sample (TLPC) ………..………...116
Table C.17 Stata Output for Random Effects Regression using Whole Sample (TLPC) ………..………...117
Table C.18 Stata Output for FGLS Regression using Whole Sample
(TLPC)…………..………..…..118 Table C.19 Stata Output for FGLS Regression with Lagged Dependent Variable using Whole Sample (TLPC)………..….119
Table C.20 Stata Output for Arellano-Bond GMM Regression using Whole Sample (TLPC) ………..………..…120
Table C.21 Stata Output for Between Estimator Regression using Shrunk
Sample (TLPC) ………..………..…121 Table C.22 Stata Output for Fixed Effects Regression using Shrunk Sample (TLPC) ………..………...122
Table C.23 Stata Output for Random Effects Regression using Shrunk Sample (TLPC) ………..………123
Table C.24 Stata Output for FGLS Regression using Shrunk Sample
(TLPC)………..124 Table C.25 Stata Output for FGLS Regression with Lagged Dependent Variable using Shrunk Sample (TLPC)………..………125
Table C.26 Stata Output for Arellano-Bond GMM Regression using Shrunk Sample (TLPC) ………..……….126
LIST OF FIGURES
Figure 3.1 Histogram of TLR……….30 Figure 3.2 Histogram of TLPC………31 Figure 3.3 Histogram of TLR for the Shrunk Sample……….31
LIST OF SYMBOLS AND ABBREVIATIONS
$ A BEDAŞ BOTAŞ ÇEAŞ DEP DPT EDAŞ EMRA EÜAŞ FGLS GCPC GDP GMM HADEP I
KCETAŞ l
LR MWh OECD OİB OLS R TEAŞ TEDAŞ TEİAŞ TEK TETAŞ TL
United States Dollars
Cross Sectional Area of Lines Boğaziçi Electricity Company Pipeline Transportation Company Çukurova Electricity Company Democracy Party
State Planning Organization Electricity Distribution Company Energy Market Regulatory Authority Electricity Generation Company Feasible Generalized Least Square Gross Consumption per Capita Gross Domestic Product
Generalized Methods of Moments Democracy Party of People Electricity Current of Lines
Kayseri and Surroundings Electricity Company Length of Lines
Technical Electricity Loss Ratio Megawatt-hour
Organization for Economic Cooperation and Development Privatization Authority
Ordinary Least Squared Resistance of Lines
Turkish Electricity Transmission-Generation Corporation Turkish Electricity Distribution Company
Turkish Electricity Transmission Company Turkish Electricity Authority
Turkish Electricity Trading and Contracting Company Electricity Theft Ratio
TOR TR V
Transfer of Operating Rights Turkish Lira
Voltage of Lines
VIF Variance Inflation Factor
THEFT AND LOSSES IN TURKISH ELECTRICITY SECTOR: EMPIRICAL ANALYSIS AND IMPLICATIONS FOR TARIFF DESIGN
by
Eray Gümüşdere
Submitted to the Graduate School of Arts and Social Sciences in partial fulfillment of
the requirements for the degree of Master of Arts
in Economics
Sabanci University Summer 2004
THEFT AND LOSSES IN TURKISH ELECTRICITY SECTOR: EMPIRICAL ANALYSIS AND IMPLICATIONS FOR TARIFF DESIGN
APPROVED BY:
Assoc. Prof. Dr. İzak Atiyas ....……….
(Thesis Supervisor)
Assoc. Prof. Dr. Alpay Filiztekin ……….
Assoc. Prof. Dr. Yıldız Arıkan ……….
DATE OF APPROVAL: ……….
© Eray Gümüşdere 2004
All Rights Reserved
ABSTRACT
THEFT AND LOSSES IN TURKISH ELECTRICITY SECTOR: EMPIRICAL ANALYSIS AND IMPLICATIONS FOR TARIFF DESIGN
Since the 1980s, electricity industries in many countries have been undergoing privatization and liberalization in order to generate improvements in efficiency and quality. Liberalization and privatization can only be successful if privatized utilities can be financially viable on their own and not rely on government subsidies except for those that are designed for social objectives such as universal service. In the case of Turkey, an important impediment to privatization and liberalization are thefts and losses that generate large asymmetries in costs across different regions. In international comparisons these costs are exceptionally high and they threaten the economic and political feasibility of reform.
This master thesis undertakes an econometric analysis to identify factors that cause such large variances in electricity theft-losses across different regions of Turkey.
Once these factors are identified, empirical results are then used to derive implications for both privatizations as well as principles that should guide tariff design, including possible subsidy mechanisms.
ÖZET
TÜRK ELEKTRİK SEKTÖRÜNDE KAYIP VE KAÇAKLAR: AMPİRİK ANALİZ VE TARİFE TASARIMINA ETKİSİ
1980’lerden bu yana elektrik sektörü bir çok ülkede verimlilik ve gelişim sağlamak için özelleştirme ve liberalleşme sürecine girmiştir. Liberalleşme ve özelleştirme ancak özelleştirilmiş kurumlar merkezi hükümetten herhangi bir finansal destek almadan kendi ayakları üzerinde durabildiği zaman başarılı olabilir. Bunun tek istisnası sadece evrensel servis gibi sosyal amaçlar olabilir. Türkiye için elektrik sektöründeki serbestleşme ve özelleştirme karşısındaki en önemli engellerden biri bölgeler arasında büyük maliyet asimetrileri yaratan kayıp ve kaçaklardır. Diğer dünya devletlerine göre kayıp kaçak maliyetleri aşırı yüksektir ve bu da yapılmak istenen reformun ekonomik ve siyasi olabilirliğini tehdit etmektedir.
Bu yükseklisans tezinde bölgeler arasıdaki muazzam kayıp ve kaçak farklılıklarına neden olan faktörlerin tanımlanması için ekonometrik bir çalışma yapılmıştır. Bu faktörler tanımlandıktan sonra, ampirik çalışma sonuçlarının özelleştirme ve tarife tasarımına iması subvansiyon mekhanizması dahil olmak üzere değerlendirilmiştir.
TABLE OF CONTENTS
ABSTRACT...iv
ÖZET...v
LIST OF TABLES...viii
LIST OF FIGURES...xii
LIST OF SYMBOLS AND ABBREVIATIONS...xiii
1. INTRODUCTION AND BACKGROUND………...1
1.1. Introduction………...……...1
1.2. Process of Liberalization in the Electricity Sector………...…………2
1.3. Distribution Private Participation………...3
1.4. The Electricity Market Law……….…….5
1.4.1. Vertical Separation………...……...………6
1.4.2. Licensing……….6
1.4.3. Regulatory Authority………..………6
1.4.4. Competition in the Generation and Retail Segment……….……..6
1.4.5. Privatization………7
2. PROBLEM DESCRIPTION……….8
2.1. Definition of the Problem……….8
2.1.2. Technical Losses………..………...9
2.1.3. Theft………...………..……..11
2.1.4. Measuring Theft-losses………..………13
2.1.5. The Size of the Problem………13
2.2. Current Tariffs and Subsidy Mechanism………15
2.2.1. The Concept of Tariff………15
2.2.2. The Current Financial Flow and Cross Subsidy Mechanism…………16
3. EMPIRICAL ANALYSIS……….………..…19
3.1. The Model………...…19
3.2. Independent Variables………19
3.2.1. Economic Variables………...20
3.2.2. Social and Cultural Variables………..………..…21
3.2.3. Variables Reflecting the Enforcement Capacity and the Reach of the State ………....22 3.2.4. Distribution Utility’s Managerial variables………...………23 3.2.5. Physical Variables………...………..…23 3.2.6. Dummy Variables……….…..………...24 3.3. The Data Set………..…….………24 3.4. Descriptive Statistics………..………....28 3.4.1. Dependent Variables………..……28 3.4.2. Independent Variables………...…32 3.5. Empirical Results………....……34 3.5.1. Year 2000 regression……….…36 3.5.2. Panel Data Analysis………...39 3.5.2.1. Multicollinearity Diagnosis………...…39 3.5.2.2. TLR Regression for the Whole Sample………41 3.5.2.3. TLR Regression for the Shrunk Sample………...…44 3.5.2.4. TLPC Regression for the Whole Sample………...…46 3.5.2.5. TLPC Regression for the Shrunk Sample……….……47 3.6. Summary and Inference………..49 4. IMPLICATIONS FOR PRIVATIZATION AND TARIFF DESIGN……....51 4.1. Privatization Concern………...51 4.2. Tariff Design Concern………54 5. CONCLUSION………...……59 APPENDIX A: THE DATA SET………..……….62 APPENDIX B: MEAN AND VARIANCE COMPARISON TESTS…………87 APPENDIX C: STATA OUTPUTS OF REGRESSIONS………..….100 BIBLIOGRAPHY………..………...127
LIST OF TABLES
Table 2.1 Cumulative Amount of Theft-Losses in MWh for 8 years) (1994- 2001………..………..…...14
Table-2.2 Methods for Tariff Regulation for Different Electricity Activities and Services………..………...….18
Table 3.1 Descriptive Statistics of Dependent Variables for the Whole
Sample………..……….….29 Table 3.2 Descriptive Statistics of Dependent Variables for the Shrunk
Sample………..……….…...29 Table 3.3 Descriptive Statistics of Independent Variables for the Whole
Sample………..………...33 Table 3.4 Descriptive Statistics of Independent Variables for the Shrunk
Sample………..………..………...…34 Table 3.5 Summary Results for Year 2000 Regressions………37 Table 3.6 Variance Inflation Factors for Year 2000 Independent Variables…..38 Table 3.7 Correlation Matrix for Independent Variables………40 Table 3.8 Principal Components for Independent Variables………..41 Table-3.9 Summary of TLR Regressions Results for the Whole Sample…..…42 Table 3.10 Summary of TLR Regressions Results for the Shrunk Sample..…..44 Table 3.11 Summary of TLPC Regressions Results for the Whole Sample..…47 Table 3.12 Summary of TLPC Regressions Results for the Shrunk Sample…..48 Table A.1 the Panel Data………63 Table A.2 Additional Data for Year 2000 Regression………..…………..84 Table B.1 Testing the Equality of Means of GDPCR for Whole and Shrunk Sample………..………...88
Table B.2 Testing the Equality of Means of RESECR for Whole and Shrunk Sample………..………...……..….88
Table B.3 Testing the Equality of Means of INDECR for Whole and Shrunk Sample………..………...89
Table B.4 Testing the Equality of Means of DEPOSITR for Whole and Shrunk Sample………..………...….89
Table B.5 Testing the Equality of Means of DEPR for Whole and Shrunk
Sample………..………..…....90 Table B.6 Testing the Equality of Means of PIE for Whole and Shrunk
Sample………..………...90 Table B.7 Testing the Equality of Means of TRRR for Whole and Shrunk Sample………..………...91
Table B.8 Testing the Equality of Means of TRPGDP for Whole and Shrunk Sample………..………...91
Table B.9 Testing the Equality of Means of PNPS for Whole and Shrunk
Sample………..………...92 Table B.10 Testing the Equality of Means of TUR for Whole and Shrunk
Sample………..………...92 Table B.11 Testing the Equality of Means of LWLPS for Whole and Shrunk Sample………..………...93
Table B.12 Testing the Equality of Means of AGRGDPR for Whole and Shrunk Sample………..……….….93 Table B.13 Testing the Equality of Variances of GDPCR for Whole and Shrunk Sample………..……….….94
Table B.14 Testing the Equality of Variances of RESECR for Whole and Shrunk Sample………..……….…94
Table B.15 Testing the Equality of Variances of INDECR for Whole and Shrunk Sample………..………..…95
Table B.16 Testing the Equality of Variances of DEPOSITR for Whole and Shrunk Sample………..………..…95
Table B.17 Testing the Equality of Variances of DEPR for Whole and Shrunk Sample………..………..…96
Table B.18 Testing the Equality of Variances of PIE for Whole and Shrunk Sample………..………..96
Table B.19 Testing the Equality of Variances of TRRR for Whole and Shrunk Sample………..……….….…97
Table B.20 Testing the Equality of Variances of TRPGDP for Whole and Shrunk Sample………..……….…97
Table B.21 Testing the Equality of Variances of PNPS for Whole and Shrunk Sample………..………..…98
Table B.22 Testing the Equality of Variances of TUR for Whole and Shrunk Sample………..………..…98
Table B.23 Testing the Equality of Variances of LWLPS for Whole and Shrunk Sample………..……….99
Table B.24 Testing the Equality of Variances of AGRGDPR for Whole and Shrunk Sample………..………..…99
Table C.1 Stata Output for Year 2000 Regression (TLR) ………101 Table C.2 Stata Output for Year 2000 Regression (TLPC) ………..…102 Table C.3 Stata Output for Between Estimator Regression using Whole Sample (TLR) ………..………...103
Table C.4 Stata Output for Fixed Effects Regression using Whole Sample (TLR)
………..……….…...104 Table C.5 Stata Output for Random Effects Regression using Whole Sample (TLR) ………..……….…105
Table C.6 Stata Output for FGLS Regression using Whole Sample (TLR)…..106 Table C.7 Stata Output for FGLS Regression with Lagged Dependent Variable using Whole Sample (TLR) ……….107
Table C.8 Stata Output for Arellano-Bond GMM Regression using Whole Sample (TLR) ………..………..……..108
Table C.9 Stata Output for Between Estimator Regression using Shrunk Sample (TLR) ………..……….109
Table C.10 Stata Output for Fixed Effects Regression using Shrunk Sample (TLR) ………..……….110
Table C.11 Stata Output for Random Effects Regression using Shrunk Sample (TLR) ………..……….111
Table C.12 Stata Output for FGLS Regression using Shrunk Sample
(TLR………..………..…….112 Table C.13 Stata Output for FGLS Regression with Lagged Dependent Variable using Shrunk Sample (TLR)……….…113
Table C.14 Stata Output for Arellano-Bond GMM Regression using Shrunk Sample (TLR) ………..………..……..114
Table C.15 Stata Output for Between Estimator Regression using Whole Sample (TLPC) ………..………...115
Table C.16 Stata Output for Fixed Effects Regression using Whole Sample (TLPC) ………..………...116
Table C.17 Stata Output for Random Effects Regression using Whole Sample (TLPC) ………..………...117
Table C.18 Stata Output for FGLS Regression using Whole Sample
(TLPC)…………..………..…..118 Table C.19 Stata Output for FGLS Regression with Lagged Dependent Variable using Whole Sample (TLPC)………..….119
Table C.20 Stata Output for Arellano-Bond GMM Regression using Whole Sample (TLPC) ………..………..…120
Table C.21 Stata Output for Between Estimator Regression using Shrunk
Sample (TLPC) ………..………..…121 Table C.22 Stata Output for Fixed Effects Regression using Shrunk Sample (TLPC) ………..………...122
Table C.23 Stata Output for Random Effects Regression using Shrunk Sample (TLPC) ………..………123
Table C.24 Stata Output for FGLS Regression using Shrunk Sample
(TLPC)………..124 Table C.25 Stata Output for FGLS Regression with Lagged Dependent Variable using Shrunk Sample (TLPC)………..………125
Table C.26 Stata Output for Arellano-Bond GMM Regression using Shrunk Sample (TLPC) ………..……….126
LIST OF FIGURES
Figure 3.1 Histogram of TLR……….30 Figure 3.2 Histogram of TLPC………31 Figure 3.3 Histogram of TLR for the Shrunk Sample……….31
LIST OF SYMBOLS AND ABBREVIATIONS
$ A BEDAŞ BOTAŞ ÇEAŞ DEP DPT EDAŞ EMRA EÜAŞ FGLS GCPC GDP GMM HADEP I
KCETAŞ l
LR MWh OECD OİB OLS R TEAŞ TEDAŞ TEİAŞ TEK TETAŞ TL
United States Dollars
Cross Sectional Area of Lines Boğaziçi Electricity Company Pipeline Transportation Company Çukurova Electricity Company Democracy Party
State Planning Organization Electricity Distribution Company Energy Market Regulatory Authority Electricity Generation Company Feasible Generalized Least Square Gross Consumption per Capita Gross Domestic Product
Generalized Methods of Moments Democracy Party of People Electricity Current of Lines
Kayseri and Surroundings Electricity Company Length of Lines
Technical Electricity Loss Ratio Megawatt-hour
Organization for Economic Cooperation and Development Privatization Authority
Ordinary Least Squared Resistance of Lines
Turkish Electricity Transmission-Generation Corporation Turkish Electricity Distribution Company
Turkish Electricity Transmission Company Turkish Electricity Authority
Turkish Electricity Trading and Contracting Company Electricity Theft Ratio
TOR TR V
Transfer of Operating Rights Turkish Lira
Voltage of Lines
VIF Variance Inflation Factor
1. INTRODUCTION AND BACKGROUND
1.1.Introduction
Until the last decades, natural monopolies all over the world have been tightly regulated. One of the most vital sectors which has been operating as a natural monopoly is the electricity industry. Nevertheless, the natural monopoly property of electricity sector has been diminishing especially in generation and distribution-retail segment due to technological and methodological improvements leading to the liberalization of these segments.1
Turkey had also tried to adapt to this trend of liberalization although the process started a bit late relative to developed countries. However, there are some problems in the different segments of the electricity sector that can act as serious barriers to the implementation of the liberalization and privatization process. Stranded costs of power plants are among such problems in generation segment, whereas high theft and losses are serious barriers to private participation in the distribution and retail segments.2 Because such asymmetric costs can harm competition and breaks down the financial viability of the system, they should be considered as primary concerns in transition to liberalization in the electricity sector.
Having mentioned two different segments, generation and distribution, the focus of this thesis will be on the distribution-retail segments since it is buyer side of the electricity sector. If this segment cannot reach competition and financial viability, none of the liberalization objectives will be achieved. The aim of this thesis is to identify factors that cause theft and losses and make policy suggestions.
1 Despite all developments in technology and methodology, transmission segment of the sector still exhibits properties of natural monopoly. In fact, distribution without retailing also shows natural monopoly feature but virtual competition can be created between distribution companies via some methods such as Yardstick Competition (Shleifer, 1985). Since in the near future distribution companies will also be retailing companies, the term “competition in distribution-retail segment” can be used conveniently.
2 Intended meaning of “ theft and losses” is “electricity theft and technical losses”
Regarding this aim, the thesis consists of two core parts other than the introduction and conclusion. The first core part presents an empirical analysis of theft and losses, intended to explain reasons and factors that cause huge variability in time dimension and especially across cross sections (among provinces). The second part uses the results of the first part and makes recommendations for privatization and tariff design so as to reduce the burden of theft-losses in the privatization process.
1.2. Process of Liberalization in the Electricity Sector
The liberalization process was initiated in 1984, by the Law No: 3096 which abolished the monopoly of Turkish Electricity Authority (TEK) and private entities participate in generation, transmission and distribution activities. Also, TEK became a state owned enterprise.
By the end of 1980s, Çukurova Electricity Company and Kepez Electricity Company which existed before TEK were given rights to do generation, transmission distribution and trading of electricity within Adana and Antalya region respectively.
By the Decree in force Law No.513 and dated 13.8.1993, TEK was slated for privatization, without cutting its relations with the Ministry of Energy and Natural Resources. As an extension of this arrangement, TEK was divided into two separate State Owned Enterprises, namely Turkish Electricity Generation Transmission Company (TEAŞ) and Turkish Electricity Distribution Company (TEDAŞ) by the Act of the Council of Ministers. 3
The most important step in the restructuring of the energy sector was the enactment of the Electricity Market Law No: 4628, (Official Gazette, 3 March, 2001).
The purpose of the law was the establishment of a stable, transparent and financially strong electricity market that works under competitive environment and private law, in order to ensure sufficient, high quality, sustainable, environmentally desireable and low cost electricity activities. (TEİAŞ web site)
This Law covers the generation, transmission, distribution, wholesale, retail, wholesale service, retail service of electricity and its import-export, together with the rights and responsibilities of all real and legal persons related with those services.
Moreover, the establishment of Energy Market Regulatory Authority and its working
agenda, principles as well as the procedure for the sector’s privatization are covered in the Electricity Market Law.
By the Decree of Council of Ministers No:2001/2026 and dated 05.02.2001 which was issued in the Official Gazette dated March 2, 2001, TEAŞ was divided to form 3 state owned public enterprises; Turkish Electricity Generation Company (EÜAŞ) Turkish Electricity Transmission Company (TEİAŞ), and Turkish Electricity Trading and Contracting Company (TETAŞ).
These three companies started to function in October 1, 2001. EUAŞ is responsible for operating and maintaining state owned thermal and hydraulic power plants and also building new plants, TEİAŞ’ responsibilities are to build and operate the transmission facilities to carry out the load dispatch, to install a communication infrastructure appropriate for the new market mechanism and to perform the balancing and reconciliation activities of the market. As for TETAŞ, it executes wholesale activity generally, buys electricity from generators (EUAŞ and others) and sells to distributors and free customers. Since our focus is on distribution segment, specific efforts in this segment should also be stated.
1.3. Private Participation in Distribution
Privatization program in the distribution sector had been started by assignment of Kayseri and Surroundings Electricity Company for Kayseri and its surroundings and also Aktaş Electricity Company for Asia part of Istanbul. Based on the Law No: 3096 (legislated in 1984), concession contracts were signed with Kayseri and Surroundings Electricity Company and Aktaş Electricity Company on March 1, 1990 and September 1, 1990 respectively. Also, both companies started to run after the physical transferring operations at the same year. (DPT, 2001)
According to the contracts, the following system has been applied for both firms.
In this system, at the beginning of each year a temporary budget is made and a buying price is determined. At the end of the year, all revenues of the firm are summed up as income and also expenses for operating the firm (which expenses are necessary and which are not are determined by the government) are recorded as expenditures. Planned investments and a reasonable dividend are also included in the expenditures. Then, the difference between income and expenditures (income minus expenses) is calculated and compared to the electricity invoices that are paid by the distribution company to TEAŞ
(TETAŞ after 2001). If the difference is greater than the invoices, the difference becomes claim of TEAŞ (TETAŞ after 2001) but if it is less, the difference becomes claim of the corresponding firm. This is actually a cost plus regulation (www.kcetas.com.tr).
However, the contracts contain no upper bounds on operating and investment expenses. Hence, there have been some conflicts between the corresponding sides and these conflicts have led to court cases.
In addition to this, distribution system was divided to 29 mission region by the Decree of Council of Ministers which was issued in the Official Gazette dated November 24, 1996 and no: 22827 (DPT, 2001).
The number 29 was increased to 33 (two of them had been already operating by Aktas and Kayseri private companies) following a decree in November 2000. (OECD, 2002) After a tender for transferring operating rights (TOR) for these corresponding areas 26 proposals were accepted and for 5 regions no proposal were offered. Among the 26 proposals 4 were cancelled by the Council of the State. Of the 22 remaining regions, 11 of them were assigned to the corresponding winners of the tender in January 1998 and franchising contracts were signed in 1999. As for the other 11 regions, evaluation and feasibility process continued. On the other hand, all these processes were further complicated in January 2000. A new law allowed TORs to be based on private law as well as public law. Some companies preferred private law, whereas other, remained under public law. After all these, the economic crisis of 2001 prompted the Treasury to be unwilling to provide guarantees for contractual obligations of TEAŞ, further jeopardizing the TOR process. Finally, transferring process was delayed.
Actually, no transfer completed yet. (OECD, 2002)
Besides, Council of State (Daniştay) cancelled the franchising contracts of Aktaş Electricity Company on February 16, 2001 claiming absence of public benefits in the franchising contract.
Moreover, on June 12, 2003 Ministry of Energy and Natural Resources seized Çukurova Electricity Company (ÇEAŞ) and Kepez Electricity Company (Kepez) by abolishing the concession agreement. ÇEAŞ and Kepez were not only electricity distribution companies but also they were involved in transmission and generation activities. In fact, their distribution activities were restricted to just definite industry consumers. Therefore, they will not be considered as distribution companies.
As for today, the only private electricity distribution company is Kayseri and Surroundings Electricity Company that has been operating since 1990 without any interruption. Now, except for this special case, all of the rest of electricity distribution activities are carried out by the state enterprises. These enterprises are TEDAŞ and its joint partnerships. TEDAŞ consists of 64+3 electricity distribution establishments (64 province establishment + substitution for Aktaş, ÇEAŞ and Kepez). Its joint partnerships are Trakya EDAŞ (Electricity Distribution Company), Boğaziçi EDAŞ, Körfez EDAŞ, Karaelmas EDAŞ, Meram EDAŞ, Sakarya EDAŞ and Başkent EDAŞ (TEDAŞ web site).
In addition to all these historical background and process, legal structure is another important aspect of the problem. Current legal structure of electricity market is based on Electricity Market Law 2001.
1.4.The Electricity Market Law
On 20th February 2001, the Turkish parliament accepted the electricity market law (Law no.4628) and it was issued on the Official Gazette dated 3rd March, 2001. The first article of the law describes the aim and the scope of the law. The first article states the purpose as establishment of financially strong, stable and transparent electricity market that operates under competitive environment and private law in order to make sure sufficient, high quality, sustainable, low cost and environmentally desiareable electricity activities.
As for scope, the law covers generation, transmission, distribution, wholesale, retail sale, retail sale service, import and export of electricity; responsibilities and rights of all entities that engage in these activities; installation of Electricity Market Regulatory Authority and its working procedure and principles and procedure for privatization of electricity entities.
This law is actually a cornerstone for Turkish electricity sector and it was written in the light of developed western countries’ electricity laws. The most important innovations introduced by the law are as follows:
1.4.1. Vertical Separation
The first unbundling in the sector was in 1993 by separation of TEK into TEAŞ and TEDAŞ. By the new law, TEAŞ was also disintegrated to three state enterprises EÜAŞ, TEİAŞ and TETAŞ which are responsible for generation, transmission and wholesale activities respectively.
1.4.2. Licensing
All public and private participants need to obtain licenses in order to engage in electricity activities. For each defined activity participants must have separate license.
Licenses are given by the central regulator, EMRA. In each license, duration of the license, price setting mechanism and license canceling conditions are attached to the license.
1.4.3. Regulatory Authority
In order to protect the sector especially from political influence and to ensure transparency, an independent and financially and administratively autonomous regulatory authority EMRA (Energy Market Regulatory Authority) has been established. (Özkıvrak, forthcoming)
This regulator is responsible for granting licenses, regulating present franchising contracts according to the law, observing the market performance, writing instructions for customer services and inspecting their applications, determining pricing principles that are mentioned in the law and making necessary adjustments, and maintaining application of the law.
1.4.4. Competition in the Generation and Retail Segment
The main focus of the law is introducing competition into the electricity sector wherever possible. In light of experience of western countries, competition for generation and retail sale was proposed. Because of market power concern, some limitations were placed on private generation firms. A private company’s total market share cannot exceed 20% of total installed capacity in the preceding year in Turkey.
As for retail sale, in addition to retail sales companies, distribution companies may also engage in retail sale by getting a retail license. Nevertheless, although retailers can
do retailing in all regions of Turkey, distributors are restricted to do retail sales just within a specified region that is identifed in their retail licenses.
Distribution companies can also engage in generation activities if they have generation licenses. However, the electricity they generate cannot exceed 20% of the electricity consumption of the previous year for their region.
TEDAŞ performs both distribution and retail sale activities until the market develops. However, TEDAŞ and its affiliates are organized as defined in the new market model and keep separate accounts for distribution and retail service activities.
1.4.5. Privatization
Ministry of Energy and Natural Resources offers its opinions and proposals to the privatization Authority (OİB) about privatization of TEDAŞ, EÜAŞ, their establishments, joint partnerships and units. Privatization procedure is carried out by OİB. In the frame of the privatization process, none of participants can have significant shares in any of the activities to control the market (Özkıvrak, forthcoming)
2. PROBLEM DESCRIPTION
2.1. Definition of the Problem
Private participation in the distribution and retailing of electricity is expected to bring substantial improvements to the sector such that investment burden of the indebted state will be reduced and inefficiency in the sector will be minimized.
However, to achieve these aims, the problem of high theft-losses needs to be overcome.
Up to now, these high ratios and high heterogeneity across provinces have been hidden by cross subsidies. That is, people in cities with low theft-losses paid some part of the bills of people who live in cities with high theft-losses. However, when private participation is realized all distribution firms will have different costs and different theft-loss ratios in different regions. Cross subsidy cannot be maintained on anymore with private participation. Moreover, if cross subsidy is abolished, due to high variability in theft-loss ratios across regions, private distribution firms will have to apply very different prices across regions. Certainly, this runs into conflict with social objectives of the state. As a result, the cross sectional asymmetry and high ratios in theft-losses prevent the establishment of financially viable competitive market in electricity distribution. Hence, factors and reasons that drive theft and losses need to be determined, it is hoped that understanding these factors will facilitate dealing with them by designing proper tariff mechanism.
In this respect, it is necessary to distinguish between losses as theft and technical losses as they are likely to be driven by different factors. Technical loss is power lost when electricity passes through transformers and lines while electricity is being transmitted and distributed. These unavoidable losses can be minimized but never eliminated. As for theft, even though they are more than technical losses in volume, they are not inevitable. However, in practice 1 or 2 percent of theft is generally seen as acceptable. While technical loss is a physical design and construction problem that must be alleviated by engineers, theft is a social problem that should be examined by social scientists. Technical losses and theft can be summed and defined as theft-loss which is unbilled electricity.
Theft-losses constitute 20% of total cost of electricity in Turkey which reaches approximately $2 billion amount annually (roughly 1% of Turkish GDP).
2.1.2. Technical Losses
Electricity is a kind of energy and according to the energy conservation law any type of energy cannot be destroyed but it can be converted to other types. Actually, this law drives the events that occur in generation, transmission and distribution of electricity.
In generation, mechanical energy is converted to electricity by means of an alternator. However, although the aim is to produce electricity, all mechanical energy cannot be totally converted to the electricity. Instead, some of mechanical power heats up the environment. That is, it is converted to partially electricity and partially to heat.
The ratio of these depends on the efficiency of the alternator and it should be noted that it is impossible to make a 100% conversion from mechanical power to electrical power.
Even though the loss due to heating up in generation is a very large amount, this is not electricity loss but total energy loss (which is also an enormous waste).4 Thus, this component of total losses will not be considered in the technical losses of transporting electricity to the consumers.
Electricity is produced at any location must be transported to the consumers. Like all other goods, transportation is costly and it is needed to construct some roads (lines) for electricity. However, in addition to the fixed cost of constructing and maintaining such roads, electricity does not need carrying activity. It carries itself by nature so it may seem to have no variable transportation cost. Actually, it does. While electrons flow into the lines they face some resistance in the wire (line). The wire absorbs some energy of the electrons and temperature of the wire increases due to the conversion of electricity to the heat. The amount of heat which appears or in other words amount of electricity loss ∆P depends on number of electrons that flow through the cross-section of the line per unit time, cross-section area of the line, length of the line, type of the wire and temperature of the line.
Formally, current (I) is defined as electrical charge quantity that passes through a cross-section of the wire per second which is a measure of flowing electrons per unit time. As for resistance of the wire (R) at a given temperature, which depends on the other variables of effecting power loss, is defined as resistivity of the wire material (ρ) times length of the line (l) divided by cross-sectional area of the wire (A). As a result,
4 For example, 1 kg coal has much more energy than electricity generated by burning of it. The difference is lost through the increase in the temperature of the environment.
power loss is equal to resistance multiplied by the squares of the current. ∆P=I²R where R= ρl/A
Resistance at a given temperature is an exogenous variable of the network that does not depend on activities of the consumers. It is a characteristic property of the physical system. Nevertheless, Current is highly sensitive to the action of network agents (consumers). By action of agents, it is meant how much power they are willing to consume at a given time. Now, it is needed to introduce a new concept, voltage, which is defined as the energy of unit charge. Hence, power is equal to current times voltage, P=VI. However, this is valid for direct current. For alternating current a new factor called power factor (cosφ) enters the equation and it becomes P=VI cosφ.
In order to understand what cosφ is, the concept of inductance, capacitance, reactance and impedance should be known. In addition to resistive property of the devices in the system, there are also two opposing features inductance and capacitance of these tools because of alternation of the current. These factors instantaneously withdraw some power; they release them to the system just a little time later and this process goes on continuously by nature. So they show some reactive property against current. This reactive property is called reactance (inductive reactance minus capacitive reactance) and it is directionally perpendicular to the resistive property (resistance).
Finally, directional resultant of these factors is called impedance. Here, φ is the angle between impedance and resistance and cosines of this angle is the ratio of resistance to impedance.
Since voltage is fixed for all consumers, if they wish to withdraw more power from the network it is realized by increase in withdrawing of current. Therefore, we can write I=P/Vcosφ where V is constant and 220 volt for low voltage consumers. As seen in the formula, current I depends on power withdrawal and cosφ therefore reactance of the devices of the consumers. When we make necessary substitutions power loss equation becomes ∆P=I²R= (P/Vcosφ)²R= (P/Vcosφ)²ρl/A.
In conclusion, if we wish to draw more electrical power from a line, power loss increase will be quadratic. Everyday, new consumers join the electricity network and they increase power load of common lines hence power loss in the lines rises quadraticly. Therefore, in order to restrict power loss, new investments should be made to construct new lines. In fact, new line construction affects the total power loss of the network by two opposite ways. First, since new lines increase the length of the network total resistance of the system increases, so does power loss. Second, however, new lines
divide the power load of existing lines and thus power loss decreases in quadratic motion driven by the equation generated above. Because the first reason of increase is linear and the second reason of decrease is quadratic, decrease effect overweighs the increase one. As a result, technical losses in the network are highly sensitive to the investments.
In addition to the losses in lines, some other losses also occur when the voltage changes. In fact, the aim of changing the voltage is to reduce total electricity losses.
This is done by transporting electricity in high voltage lines where losses are low5 and when electricity is close to the target, its voltage is reduced for consumption. This reduction of voltage is also costly in terms of electricity loss, but this loss amount is much lower than what it would be if electricity was totally be transported in low voltage lines. The voltage reduction process is realized by transformers and some losses occur in transformers driven by two effects. Some losses are function of voltage and some others are function of current. Because voltage is fixed, only current and therefore consumption of consumers affects the power loss in transformers. These transformers have some capacity and if power and therefore current drawn increases, their utilization increases and power losses rise, too. If power consumption is too high, reaching the capacity of transformer, the transformer may break down.6 Such overloads increase number of power outages and reduce the quality of electricity such as voltage stability.
Therefore, transformer utilization ratio may be a good indicator for power losses and investments. Since high utilization ratio means insufficiency in setting up required transformers, lack of investment reveals itself.
As a result of all these considerations above, 3 variables which are investments, transformer utilization ratio and low voltage line lengths will also be used in regression analysis below to capture some of the variability in technical losses.
2.1.3. Theft
In addition to the technical losses, another reason for energy loss is theft. This is actually selling losses and results from not technical but social reasons. Smith, 2003
5 When voltage is high, current is low since multiplication of current and voltage is constant. Therefore, power loss in high voltage lines is less.
6 Prof. Özay from Middle East Technical University says hundreds of transformers are broken down in Istanbul, annually.
suggests 4 types of electricity theft that differ in terms of methods used to steal electricity.
Fraud: In this type of theft, consumer tries to cheat the utility. Most general version of this is meter tampering. The consumer intervenes in the meter to make it show amount consumed less than it actually is. Actually, this is risky when done by amateurs. However, in some countries Malaysia for instance, professionals who are the managers of electricity utilities do this for a moderate price (New Strait Times, 1999).
This is an organized crime, actually. When this meter tampering is done by organizations professionally, it may cause large amounts of losses.
Steal: This method can be considered as the most direct method to thieve. In contrast to the other 3 methods, the consumer does not pay anything for electricity. The consumer does not have any relation with electricity utility in this method. The consumer fixes a line to the electricity grid and draws power via this line. This method is especially quite common in poor residential areas. In fact illegal lines are easy to detect but police force is needed to remove those lines (Smith, 2003). Bribery is also another problem with this method that officers may condone stealing electricity when they accept bribe.
Billing irregularities: In some cases, the consumer may pay less or more than he consumes because of billing irregularities. The utility may be unable to detect the consumption correctly, leading to less or more amounts in billing than correct consumption. One of the most general reasons of billing irregularities is bribe. The consumer may bribe to pay less and meter reader officer may gain unofficial payment.
(Smith, 2003)
Unpaid Bills: Although Smith (2003) suggests this as a type of electricity theft, we do not consider unpaid bills as theft. Actually, it is a revenue collection problem rather than theft. Nevertheless, it is also some unpaid amount which causes financial weakness to the electricity utility and it should be overcome. In our analysis this amount is not included in theft amounts.
After explaining sources of thefts, it is useful to show general picture of theft- losses in the world. When theft-losses are inspected for years of 1980 and 2000 over 102 countries using the World Bank data, the results show that except a little decline in Western Europe, North America and Australasia, theft-losses have raised dramatically (Smith, 2003). Considering the technological improvement which causes reduction of technical losses, it can be deduced that theft is the most effective factor driving theft-
losses to high levels. Hence, electricity theft is getting more dangerous for most of the countries and threatening the financial viability of electricity sector.
2.1.4. Measuring Theft-losses
As for measuring theft and losses, some different indicators can be used.
Nevertheless, there is no simple way to decompose theft and losses; what is available as data is the sum total of theft and losses which is calculated by subtracting billed consumption from total energy given to the grid. Summed amount of theft and losses which referred to amount of theft-losses is the first measure used in this analysis to see the total cost and size of the problem. However, like gross domestic product (gdp) this is not a scaled measure and is highly sensitive to population and consumption level of the corresponding region. As division of gdp by population gives more accurate and reasonable measure for income level, theft-losses amount can also be scaled by population which gives theft-loss per capita TLPC. Moreover, a better scaling factor is consumption rather than population since it precisely gives percentage cost of theft- losses. The resulting measure is called theft-loss ratio TLR. Both TLR and TLPC are going to be used in empirical analysis and it will be checked whether factors effecting TLR and TLPC significantly are different or not.
2.1.5. The Size of the Problem
Now, it could be a good idea to look at total amount of theft-losses. Table 2.1 shows total amount of electricity in MWh that had been lost or thieved for each province between 1994 and 2001 calculated by TEDAŞ Statistics.
As seen in the table 2.1, Istanbul, Diyarbakir, Ankara, Sanliurfa and Mardin are the leading five provinces in terms of the absolute level of theft-loss cost to Turkey.
Also, Istanbul should be underlined since its cost is approximately as much as sum of the remaining four top five cities. Hence, to combat with financial costs due to theft, primary target should be those cities mentioned and especially Istanbul. If it is assumed that price of electricity is approximately $80 per MWh, total cost of Istanbul’s theft- losses for 8 years is about $2 billion. The cost of total theft-losses to Turkey for the same period is about $15 billion. Thus, theft-losses inflict costs to Turkish economy $2
billion anually which is %1 of Turkey’s GDP. These numbers actually show how serious the problem is.
Table 2.1 Cumulative Amount of Theft-Losses in MWh for 8 years (1994-2001)
ADANA 2,507,873 EDİRNE 566,474 KÜTAHYA 412,942 ADIYAMAN 649,738 ELAZIĞ 849,018 MALATYA 954,449 AFYON 433,776 ERZİNCAN 159,483 MANİSA 1,070,922 AĞRI 768,084 ERZURUM 1,009,725 MARDİN 4,709,896 AKSARAY 255,397 ESKİŞEHİR 851,402 MUĞLA 1,240,231 AMASYA 434,509 GAZİANTEP 2,410,352 MUŞ 816,520 ANKARA 7,818,051 GİRESUN 341,651 NEVŞEHİR 315,698 ANTALYA 1,721,408 GÜMÜŞHANE 97,548 NİĞDE 336,254
ARDAHAN 222721 HAKKARİ 761574 ORDU 574,519
ARTVİN 275,851 HATAY 1,504,701 OSMANİYE 192374
AYDIN 1,139,976 IĞDIR 390190 RİZE 457,892
BALIKESİR 1,127,914 ISPARTA 354,421 SAKARYA 1,238,976 BARTIN 246,329 İÇEL 2,750,168 SAMSUN 1,618,539 BATMAN 2,334,227 İSTANBUL 24,438,259 SİİRT 896,095 BAYBURT 43,099 İZMİR 4,664,723 SİNOP 253,984 BİLECİK 217,092 K.MARAŞ 1,122,094 SİVAS 468,939 BİNGÖL 265,138 KARABÜK 195,952 ŞANLIURFA 7,458,933 BİTLİS 518,728 KARAMAN 111,366 ŞIRNAK 2512258 BOLU 541,693 KARS 744,793 TEKİRDAĞ 1305751 BURDUR 278,135 KASTAMONU 280,997 TOKAT 488,055 BURSA 3,247,540 KAYSERİ 868,276 TRABZON 860,528 ÇANAKKALE 587,195 KIRIKKALE 303,615 TUNCELİ 119,303 ÇANKIRI 128,000 KIRKLARELİ 776,998 UŞAK 339,853 ÇORUM 415,948 KIRŞEHİR 247,282 VAN 1,994,367
DENİZLİ 916,777 KİLİS 177039 YALOVA 306941
DİYARBAKIR 8,194,247 KOCAELİ 3,070,693 YOZGAT 373,998 DÜZCE 467755 KONYA 1,300,009 ZONGULDAK 921,447
Although the absolute level of the amount is important as total cost incurred by the whole country, its importance vanishes when the main concern is the effect of theft- losses on privatization. Because privatized utilities are interested in profits, unit cost of theft-loss becomes important so they should control theft-loss ratios in their regions.
Even though the absolute level of theft-loss amount in a region may be relatively high, the corresponding utility may still succeed to run the firm and make a profit if the net consumption is so high, too. For instance, Izmir’s theft-loss amount is ranked 6th in Turkey and this amount is just slightly less than Mardin’s but Izmir is one of the unproblematic cities in terms of electricity theft-losses. The reason is that, despite relative high amount in theft-loss of Izmir, its consumption is also huge, making total theft-loss cost less than 10% which is near to OECD standards. Therefore, the most
critical variable is theft-loss ratio TLR (amount of theft loss divided by gross consumption which is total electricity drawn from the grid) in terms of the financial viability of the private distribution utilities. In order to make the concept of financial viability clear, tariffs concept, financial flow and subsidy mechanism in the electricity sector also need to be understood.
2.2. Current Tariffs and Subsidy Mechanism
2.2.1. The Concept of Tariff
Tariff is a regulation of revenue and pricing among consumers, producers and other third persons. It regards all parts’ rights. A tariff consists of several components.
These components are investments, operation expenses, costs due to transferring of operating rights, cross subsidies, taxes, insurances, funds, stranded costs and regulator’s fees. (Sevaioglu, 2004) The sum of all these components gives the bill price.
Practically, these components may differ across different regions and different types of consumers, leading bill prices to differ. Actually, for purely allocative efficiency marginal costs must be borne by the consumer of the corresponding service.
Nevertheless, sometimes economic and social aims contradict in the design of tariff structure for electricity. (Dossani, 2004) Economic efficiency suggests different pricing, whereas social aims may require applying a single tariff all over the country. In this regard, up to now, there has been a single tariff for every region of Turkey, although costs of different regions dramatically differ.7 The reason of these differences is especially high variance of thefts and losses among different provinces.
In fact, different parts of a city also differ in terms of electricity thefts and losses.
By this logic, continuing the reduction of region size or increasing the number of groups that should be charged with different tariffs, eventually leads to the idea that each subscriber has own unique cost and so should has own unique bill price. However, even logic that takes each subscriber as a separate cost unit, theft costs should not be reflected to each consumer separately since theft is not an actual cost caused by the corresponding consumer rather it is a weakness of the electricity system. Thus, according to this logic as well theft must be reflected to prices uniformly all over the
7 Actually, tariffs are a little lower in provinces which have priority for development.
country. Hence, as far as reflecting the costs of theft is concerned, considering the country as a single unit gives the same result as considering each consumer as a single unit.
Logical constructions aside, it is important to consider practically feasible design of tariffs. Firstly, the current situation in Turkey is going to be explained.
2.2.2. The Current Financial Flow and Cross Subsidy Mechanism
As mentioned above, Turkey has always applied a uniform national tariff in electricity all over the country and is still applying in spite of the articles of Electricity Market Law, 2001. This law introduces competition and economic efficiency into the electricity sector so it proposes regional tariffs that reflect corresponding costs in the distribution of electricity. However, the current government has not implemented regional tariffs especially due to political reasons and pressure of southeastern deputies.
Currently, a national single tariff is in operation. The most striking property of this tariff is equality principle. It applies almost same tariffs across regions and consumption purposes (residential, agricultural or industrial usage) unlike most developed countries.
This equalization mechanism is achieved through cross subsidies, meaning that low cost consumers subsidize high cost ones causing a single final price. Actually, there are 5 types of cross subsidies in electricity pricing. These are:
1. Across subscriber groups (industrial, residential, agricultural usage etc.) 2. Across regions or provinces
3. Across vertical activities (distribution, retail, generation, wholesale etc.) 4. Across institutions of horizontal activities (hydro, wind, natural gas etc.
usage in generation.)
5. Across sectors ( e.g. from natural gas to electricity)
This classification will be helpful in explaining how the current subsidy mechanism works. In designing tariffs, one approach is to start with costs and derive the retail price as the sum off all costs. This approach gives different prices for different cross sections (e.g. provinces) since costs are different for each cross sectional unit. The other approach is to start with a final price8 and subtract each cost element, reaching at the end a “loss” or “profit” as a residual. All these residuals are then equated to zero by
8 This final single price is selected such that sum of all costs over the whole country equal to selected final price times the quantity sold.
